DIGITAL ETHICS

Digital images raise ethical issues that have so far received only little attention. The discrete pixels of digital images can be freely combined. Compared to the dissemination of analog images the net enables dis-proportionately greater control over how digital images are to be distributed. As of today, the consequences for contemporary visual communication as well as for the cultural visual memory are all but clear. Can there be an applied ethics of digital images at all? What could be the content of such a normative ethics? And what is the relationship between ethics and the law of digital images? This volume brings together the contributions to an interdisciplinary German-Italian conference at Villa Vigoni, sponsored by the DFG. 

With contributions by
Prof. Gianmaria Ajani, Prof. Tiziana Andina, Dr. Eva-Maria Bauer, Dr. Davide Dal Sasso, Prof. Dr. Thomas Dreier, PD Dr. Johannes Eichenhofer, Prof. Maurizio Ferraris, Prof. Dr. Christophe Geiger, Prof. Dr. Dr. h.c. Werner Gephart, Olivia Hägle, Prof. Wybo Houkes, Dr. Lisa Käde, Prof. Massimo Leone, Lorenz Müller-Tamm, Dr. Eberhard Ortland, Prof. Dr. Benjamin Raue, Ass. Prof. Cosetta Saba, Prof. Dr. Reinold Schmücker, Ass. Prof. Enrico Terrone and Prof. Dr. Wolfgang Ullrich.

More and more often, decisions with vast impact on our everyday life are made by artificial intelligence (AI), be it the granting of mortgages, be it in medical situations, in crime-prevention or in such mundane features as web searches.
Up until now, only weak artificial intelligence is available.It is called weak AI because it only performs well in a very specific field, e.g. in playing chess, or in the case of industrial robots, assembling cars.The so called strong AI that would be able to transfer acquired knowledge to other situations and contexts, an AI that is able to decide for itself even in uncertain or unforeseen situations and that reflects upon its own actions and decisions is still far away.But once it has evolved, it can either lead to a positive or a negative extreme, depending on how it will be handled.
This leads to the question whether it will be possible to check on the algorithms used and whether one will be able to control the software.Which consequences on politics, lawmaking and society will result from this technical development?How can we make a positive impact on these topics and how can we enable our fellow human beings to understand and shape this innovative world of tomorrow?
More and more decisions that are made by human beings are delegated to software algorithms now.This increases the complexity as well as the velocity of social life.Artificial intelligence holds new challenges for citizens, governments and companies all over the world, influencing not only algorithms but also data aggregation and network infrastructure.
Bisher wurde "lediglich" schwache künstliche Intelligenz entwickelt, die in ganz spezifischen Feldern extrem gut funktioniert: Software, die Schachspiele gewinnt; Roboter, die Autos zusammenbauen.Die Technologie ist jedoch noch nicht so weit, beispielsweise Gelerntes vom einen Kontext in einen anderen zu übertragen, unter Unsicherheit selbständig zu entscheiden, oder über sich selbst zu reflektieren.Erst wenn sie dazu in der Lage ist, sprechen wir von starker künstlicher Intelligenz, die im Zweifel in ein negatives oder positives Extrem führen könnte.It is out of question that artificial intelligence will be further developed.What we must do is to shape it according to criteria, beneficial to us humans.This is the reason why we as Deutsche Telekom have devised digital ethics guidelines on AI for our AI processes and AI products and their usage.We hereby not only set a framework for ourselves but we also want to invite other companies, consumer associations, politicians and citizens to discuss ethical questions in this area.
With our digital ethics initiative, we are at the forefront of the discussion in Germany.Taking the initiative, dealing with a topic and understanding it, helps to reduce people's fear of the unknown.To quote the physicist Marie Curie: 'One does not need to fear anything in life, one only needs to understand everything.' As human beings, it is our decision which values artificial intelligence shall mirror in the end.Each and every human generation must take responsibility for respecting these values and this must never be forgotten.

ORIGINS:
The need for digital ethics became apparent as autonomous cars came into existence, bringing up questions of how much power should be given to a machine to decide in a critical situation.During a business trip to Israel, we talked to high-tech companies, universities and startups about their views on the use of artificial intelligence to understand it better.
Artificial intelligence provides us with various possibilities and chances, each coming with ethical and societal challenges.Which ethical specifications shall a software developer follow while coding an algorithm, and who should define these specifications?
The majority of people agrees that the development of AI algorithms shall obey legal prescripts and that it shall conform to our set of values.Fair enough.But what does that mean in detail?How does, for example, 'our set of values' look like?Different cultures have different values, but which one shall now influence the development and the use of AI?
The scramble for this new technology also represents the battle between the liberal-capitalist Silicon Valley and the authoritarian-communist People's Republic of China, both representing distinct ideological systems.On grounds of the European culture and its values, Europe positions itself in this fight.In the end the one that masters the basic knowledge of AI technology, the one that develops and produces the key AI products and that determines the rules of usage will be leading the debate.Since AI as a technology itself is neutral, it is our responsibility as human beings to decide what we want to utilize AI for.Technology only serves as a tool; the rights and wrongs of its usage must be and are questioned by digital ethics.

DIE ENTSTEHUNGS-GESCHICHTE:
For Deutsche Telekom, AI already is an existing tool in the company to improve our customers' experience.For example, the customer service in Austria is being supported by our chatbot Tinka and our customer service in Germany will soon be supported by the chatbot Eliza, answering customers' questions.'Hello Magenta' is a virtual assistant complementing DT's smart speaker whereas our T-Systems colleagues use AI to enable the German railway to better predict their trains' punctuality.
As a basis for the preparation of the digital ethics guidelines on AI, we invited many parties within the company to join the discussion -internationally as well as across disciplines and departments.We also included the departments which deploy AI and develop as well as design these technical products.Experts in the departments of Technology and Innovation, the Telekom Innovation Laboratories and the Telekom Design Gallery as well as specialists in the fields of IT-security, data security, finance, customer service, procurement, partnering etc. contributed to the many workshops we conducted -every single person acting out of conviction, resulting in the guidelines for the ethical use of AI you will find in the following.
We do not presume that our guidelines are the answer to everything.But with these guidelines, we Es ist an uns, den Menschen, zu bestimmen, wofür und wie wir KI einsetzen, denn die Technologie an sich ist neutral.Sie ist nur ein Werkzeug der handelnden Personen.Digitale Ethik fragt hierbei nach dem Guten und Richtigen im Zusammenhang mit der Nutzung von künstlicher Intelligenz.
Das Ziel: Leitlinien für einen ethischen Umgang mit KI zu entwickeln.Wir maßen uns hierbei nicht an, mit unseren Leitlinien den so genannten "Stein der Weisen" gefunden zu haben, sondern wollen die want to take the discussion on ethics in times of AI even further, aiming for a broad consensus in the industry as well as society.
With this in mind, our activities are not only focused on our organization itself.We want to reach citizens outside of Deutsche Telekom just as much as we want to reach our employees.Within the frame of the 'Expertenkreis Digitale Verantwortung', we presented our AI guidelines to a representative group of the civil society on the 10th of April 2018.On this occasion, a broad group of people joined the discussion, e.g. a startup founder, a museum curator, a professor and a pupil, a Science Fiction author and head of a politics division, a federal minister of family, the elderly, women and youth, a teacher and many more.
The fruitful discussion resulted in a collection of meaningful advices that we took to heart.With this first public trial, we laid the cornerstone for a common understanding for encouraging and demanding the right usage of AI.
Within Deutsche Telekom, we nourish the concrete realization of our digital ethics guidelines: in addition to the consultation of our data security adviser as well as our board members, we have started an internal communication campaign, are setting up eLearnings for our employees, are conducting workshops to improve products for our customers, and we are developing professional ethics and AI Enterprise Architecture Principles (meaning, IT architecture for AI algorithms).In addition, we analyze processes and look at committeesjust to name a few of the internal activities.
Looking at the DT external measures, an 'Alliance for Digital Ethics', representing an interdisciplinary association of German technology companies, nationwide federations and institutes, is being established.Its aim is to encourage and promote the discourse and the sharing of views on digital ethics, and especially ethics in terms of artificial intelligence.Furthermore, we have sought the exchange of ideas between experts, scientists and companies worldwide.We talked to companies like Amazon, Google or Microsoft, Elon Musk's non-profit research company Open AI or the Partnership on AI (PAI), and we also talked to universities and startups, discussing digital ethics and presenting our guidelines on AI.
As an internationally operating company and as one of the 'World's most ethical companies 2018', Deutsche Telekom recently joined the Partnership on AI, a technology industry consortium researching and formulating best practices for AI systems, founded by companies including Apple, IBM and Microsoft.In this consortium, we want to and we will actively take part in the development of global best practices and technologies like artificial intelligence to support companies and society.Als international agierendes Unternehmen und eines der "World's most ethical companies 2018" ist die Deutsche Telekom seit kurzem auch ein Partnerunternehmen der Partnership on AI, einer gemeinnützigen Organisation auf dem Gebiet der KI, gegründet von großen IT-Unternehmen wie Apple, IBM und Microsoft.Durch diese Partnerschaft können wir nun auch aktiv an der Entwicklung globaler "best practices" und Technologien wie der künstlicher Intelligenz zum Nutzen von Unternehmen und Gesellschaft mitwirken.

DEUTSCHE TELEKOM'S GUIDELINES ON AI:
The digital ethics guidelines on AI mirror Deutsche Telekom's understanding of AI and its usage.They represent the framework for our fields of application as well as the manner of application of the AI technology.They set the rules that we need for the work of our developers and designers and for when we collaborate with suppliers and partners.The digital ethics guidelines on AI also provide legal certainty when it comes to the handling of AI products and services -i.e. by stating the demand for data privacy and data security measures in data aggregation and processing activities (Big Data) or by expecting AI algorithms used for interpretation and decision making to be flawless (bias free) and tested beforehand.
Looking at Deutsche Telekom's AI guidelines they might seem too plain and simple for such an important and far-reaching topic.To some people they do not hold any greater insights.Nevertheless, each and every guideline has to be seen as a statement and a promise: for taking on responsibility, for transparency, for data privacy, security, for self-determination in terms of data, for consumer protection.In short: they vouch for the values we know and that we expect as a user when interacting with AI products and services, but that we cannot always be sure of being considered.
Our aim is to enable people to self-reliantly use new technologies.Software coders and technicians shall responsibly develop and improve these technologies and we must provide them with an ethical beacon as a point of orientation.Furthermore, boundaries within which self-learning systems can operate must be defined and initially implemented by developers.
Wir sind davon überzeugt, dass sich das große Potential von KI nur verwirklichen lässt, wenn diese an den Bedürfnissen der Menschen ausgerichtet Due to our technical and societal expertise in terms of communication technology, it is our self-defined duty to consider ethical challenges right from the beginning, to deliver answers to these ethical questions and to check whether the answers still hold true in our ever-changing world -continuously.

Herzlichst, Marie von der Groeben und Stefan Kohn
It is about clarity, transparency, security and trust, in summary, it is about the basis for our customers.We represent the key messages of our AI guidelines as a company and in favor of our customers and their trust in us.
To give some examples: the digital ethics guidelines on AI call for a clear definition of who is responsible for which AI system or feature.We commit ourselves to making it clear to the user whether they are interacting with an AI enhanced application.This also results in our understanding of the need to provide a stable and secure basis for data and the infrastructure.Hence, risk and dependency analysis as well as the initial question of why a system or feature shall be developed and whether it caters to a customer's needs are of huge importance.All nine guidelines are illustrated on the following pages.
In a digital world that is being enhanced with AI, each and every human being must be able to self-reliantly and freely decide where and to which extent technology is to support them, and in which situations no AI based service shall help them to accomplish a task.In addition to that, an AI system's results that have an impact on the human being must be comprehensible -especially when the individual is not able to alter the results.
We believe in the huge potential that lays in artificial intelligence.And we are convinced that this potential can solely be activated and achieved if the use of AI only caters to the human beings' needs while being accepted by society.The development and use of AI must obey ethical values and legal rules, always ensuring the humans' sovereignty.

PREAMBLE:
Being a trustful partner and providing the best customer experience are two of the main goals of the Deutsche Telekom.
As one of the leading ICT companies in Europe, it is our responsibility to foster the development of smart technologies, and to emphasize the importance of defined ethical rules that these technologies as artificial intelligence shall follow.
It needs a common understanding on AI to build a correct ethical framework that shall apply to these technologies and their usage.Nowadays, there are different definitions of AI, one of the first being described in the research project proposal made by McCarthy, Minsky, Rochester and Shannon in 1956: 'Every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.'To us -as to other companies and key players in AI -AI in machines and computer systems imitates the processes and characteristics that mark human intelligence, such as being able to learn, to think logically and to correct oneself.
During recent decades, artificial intelligence has become one of the most intriguing as well as challenging topics of our time.With not only experts, politicians and the public discussing it, AI is prominently placed in our lives.Artificial intelligence starts to have an ever-bigger impact on different aspects of living.It influences the economy by boosting efficiency and generating new fields of business.AI plays a role in society when it comes to broad discussions on subjects such as autonomous cars, 'job generator vs. job killer' etc., showing the effects AI technology can have, and

Um einen entsprechenden ethischen
Nevertheless, not only does AI carry chances and possibilities but it also confronts us with ethically critical situations that never existed before.Ethical principles are based on relationships between human beings as well as traditional values -until now.In this new era in which technology has such a huge impact on our lives, we need to bring these principles up to date to be prepared for the challenges awaiting us.
Deutsche Telekom develops and deploys AI technology for the sake of simplifying our customers' lives and to improve their experience and therefore their satisfaction.Meanwhile, we as DT have initiated various projects within the company which work on and with AI.Naturally, this results in an increasing 'digital responsibility', and we must make sure that the ethically correct use of AI stays in focus.In these times, we as DT want to provide answers to our customers, partners, stakeholders and to the public.
In the light of the significant paradigm shift awaiting us, the following digital ethics guidelines on AI shall give advice on how we want to shape the future with AI.To us, technology serves one main purpose: it shall support us human beings.Therefore, AI shall extend and complement a human's abilities rather than lessen them.
Wir als Deutsche Telekom entwickeln und nutzen künstliche Intelligenz, um die mit dieser Technologie einhergehenden Vorteile, wie etwa die generelle Vereinfachung des Alltags oder die Verbesserung der Kundenzufriedenheit, zu nutzen.Derzeit ist bei uns bereits eine Vielzahl an Projekten ins Leben gerufen worden, die an und mit künstlicher Intelligenz arbeiten.Hiermit einhergehend steigt unsere "digitale Verantwortung" und wir müssen dafür sorgen, dass KI auf ethische Weise genutzt wird.Vor diesem Hintergrund möchten und werden wir als Deutsche Telekom hierauf unseren Kunden, Geschäftspartnern, Aktionären und der Öffentlichkeit Antworten geben.Die folgenden KI-Leitlinien für die Entwicklung und Nutzung von künstlicher Intelligenz sollen diesem bedeutenden gesellschaftlichen Wandel, dem wir gegenüberstehen, Rechnung tragen und darstellen, wie wir die Zukunft mit KI gestalten wollen.Für uns The human always remains responsible.Our solutions come with a clear definition of who is responsible for which AI system or feature.We are in charge of our products and services.And, we know who is in charge for partner or third party solutions.
With AI technology being in its infancy, we are aware of our responsibility in development -from the very beginning.We make sure that we clarify which initiative or product owner has which responsibilities.For partners or third parties, we define clear guidelines for when a partnership can be established.And, we declare which duties are connected to the respective AI parts.

N° 1
We are responsible.
We act in tune with our company values.Our systems and solutions must subordinate to human-defined rules and laws.Therefore, in addition to our technical requirements, our systems and solutions have to obey the rules and laws that we as Deutsche Telekom, our employees -and human beings as such -follow.
AI systems have to meet the same high technical requirements as any other IT system of ours, such as security, robustness, etc.But since AI will be (and already is) a great part of our everyday lives, even guiding us in several areas, AI systems and their usage also have to comply with our company values (Deutsche Telekom's Guiding Principles and Code of Conduct), ethical values, and societal conventions.We have to make sure of that.
We enrich and simplify our customers' lives.If an AI system or the usage of customerrelated data helps us to benefit our customers, we embrace this opportunity to meet their demands and expectations.
The aggregation and use of customer data -especially in AI systems -shall always be clear and serve a useful purpose towards our customers.Systems and processes that support in the background are as important as services that interact with our customers directly.

N° 3
We put our customers first.
In no case we hide it when the customer's counterpart is an AI.And, we are transparent about how we use customer data.As Deutsche Telekom, we always have the customer's trust in mind -trust is what we stand for.
We are acting openly to our customers.It is obvious to our customers that they are interacting with an AI when they do.In addition, we make clear, how and to which extent they can choose the way of further processing their personal data.

N° 4
We are transparent.
Data security is a prime quality of Deutsche Telekom.In order to maintain this asset, we ensure that our security measures are up to date while having a full overview of how customer related data is used and who has access to which kind of data.
We never process privacy-relevant data without legal permission.This policy applies to our AI systems just as much as it does to all of our activities.Additionally, we limit the usage to appropriate use cases and thoroughly secure our systems to obstruct external access and ensure data privacy.

N° 5
We are secure.
Our AI solutions are developed and enhanced on grounds of deep analysis and evaluation.They are transparent, auditable, fair, and fully documented.We consciously initiate the AI's development for the best possible outcome.
The essential paradigm for our AI systems' impact analysis is "privacy and security by design".This is accompanied e.g. by risks and chances scenarios or reliable disaster scenarios.We take great care in the initial algorithm of our own AI solutions to prevent so called 'Black Boxes' and to make sure that our systems shall not unintentionally harm the users.

N° 6
We set the grounds.
We are able to deactivate and stop AI systems at any time (kill switch).Additionally, we remove inappropriate data to avoid bias.We have an eye on the decisions made and the information fed to the system in order to enhance decision quality.
We take responsibility for a diverse and appropriate data input.In case of inconsistencies, we rather stop the AI system than pursue with potentially manipulated data.We are also able to "reset" our AI systems in order to remove false or biased data.By this, we install a lever to reduce (unintended) unsuitable decisions or actions to a minimum.

N° 7
We keep control.
We believe that human and machine intelligence are complementary, with each bringing its own strength to the table.While we believe in a people first approach of human-machine collaboration, we recognize, that humans can benefit from the strength of AI to unfold a potential that neither human or machine can unlock on its own.
We recognize the widespread fear, that AI enabled machines will outsmart the human intelligence.We as Deutsche Telekom think differently.We know and believe in the human strengths like inspiration, intuition, sense making and empathy.But we also recognize the strengths of AI like data recall, processing speed and analysis.By combining both, AI systems will help humans to make better decisions and accomplish objectives more effective and efficient.
We acknowledge the transformative power of AI for our society.We will support people and society in preparing for this future world.We live our digital responsibility by sharing our knowledge, pointing out the opportunities of the new technology without neglecting its risks.We will engage with our customers, other companies, policy makers, education institutions and all other stakeholders to ensure we understand their concerns and needs and can setup the right safeguards.We will engage in AI and ethics education.Hereby preparing ourselves, our colleagues and our fellow human beings for the new tasks ahead.
Many tasks that are being executed by humans now will be automated in the future.This leads to a shift in the demand of skills.Jobs will be reshaped, rather replaced by AI.While this seems certain, the minority knows what exactly AI technology is capable of achieving.
Prejudice and sciolism lead to either demonization of progress or to blind acknowledgment, both calling for educational work.We as Deutsche Telekom feel responsible to enlighten people and help society to deal with the digital shift, so that new appropriate skills can be developed and new jobs can be taken over.And we start from within -by enabling our colleagues and employees.But we are aware that this task cannot be solved by one company alone.Therefore we will engage in partnerships with other companies, offer our knowhow to policy makers and education providers to jointly tackle the challenges ahead.
General remark: the effects of AI on the job market and tasks within Deutsche Telekom (increasing added value by expanding the area of activity or enhancing efficiency) are not discussed in the following guidelines.dient Technologie einem Hauptzweck: Sie muss unterstützend wirken.Daher sollte KI auf jeden Fall die menschlichen Fähigkeiten erweitern und ergänzen, sie aber keinesfalls einschränken.Anmerkung: Die Auswirkung von KI auf Arbeitsplätze der Deutsche Telekom -sei es zur Steigerung der Wertschöpfung im Sinne einer Tätigkeitsbereicherung und -erweiterung oder auch im Sinne von Effizienzsteigerung -ist nicht Gegenstand dieser Leitlinien.